A Novel CSUA Based Data Mining Approach for Mobile Computing Environments

lowlyoutstandingMobile - Wireless

Nov 24, 2013 (3 years and 11 months ago)

70 views

International Journal of Computer Applications (0975


8887)

Volume 7


No.
4
, September 2010

A Novel CSUA Based Data Mining Approach for Mobile
Computing Environments
Ashutosh K. Dubey

Dept. of Computer Science &

Engineering

TITR,
Bhopal, India


Ganesh Raj Kushwaha

D
ept. of Computer Science &
Engineering

TITR, Bhopal, India


Jay Prakash

Dept. of Computer Science &
Engineering

IEM, Mathura, India

ABSTRACT

Data mining services play an important role in the
field of
Communication

industry.

Data mining is also called knowledge
discovery in several database including mobile databases.

In this
paper, we discuss and analyze the consumptive behavior based on
data
mining technology. We discuss and analyze different aspects of
data mining techniques and their behavior in mobile devices. We also
analyze the better method or rule of data mining services which is
more suitable for mobile devices.
In this paper, we propo
se a novel
CSUA

(Create,

Select,

Update and Alter) Based data mining
approach

for mobile computing
environments. In CSUA approach,
we first propose to create a data set according to the requirement and
need then select the data set, then update the data se
t according to the
customers and finally reconstruct the structure of the dataset.

It is
achieved by CLDC and MIDP component of J2ME.


Keywords

Data Mining,

CSUA, CLDC
,

MIDP
.



1.

INTRODUCTION

The innovations in computer science have made it possible to acq
uire
and store enormous amounts of data digitally in databases, currently
giga or terabytes in a single database and even more in the future.
Many fields and systems of human activity have become increasingly
dependent on collected, stored, and processed i
nformation. However,
the abundance of the collected data makes it laborious to find
essential information in it for a specific purpose. In the late 1980’s,
the disciplines of
knowledge discovery
and
data mining
emerged to
help survey the information conten
t of data. It

is also use in mobile
devices with

the use of MIDLET and CLDC component of J2ME.

In
few years back, mobile extensions to Grid systems have

been
increasingly proposed in order to support ubiquitous

access
and
selection
to the Grid and to inclu
de mobile devices as additional

Grid
resources [1, 2].



In today’s scenario mobile devices, such as mobile phones, PDAs,
notebook and others, provide a basic building block
[3][4][5][6].
Finding prevalent mobile user patterns and
behavior in

large amount o
f data has been one of the major problems in the area
of mobile data mining. Particularly, the algorithms of discovering
frequent user’s behavior patterns in the mobile agent system have
been studied extensively in recent years. The key feature in most of
these algorithms is that they use a dataset and frequent Item
-
Sets
visited by the
customers.
In this case, some problems occur because
they do not consider that mobile user’s
behavior

patterns are
dynamically variable as time passes. In this paper we discu
ss some of
the
data mining

service which are use in different areas and then
apply those services to mobile devices and then apply those DMS
services in mobile computing and exploiting the need of DMS in
mobile computing environments using CLDC and MIDP
co
mponents.

The Connected, Limited Device Configuration (CLDC)
and the Mobile Information Device Profile (MIDP) have emerged as
J2ME standards for mobile phone applications development which
are used with DMS services.
The role of CLDC and MIDP
component is
to apply Data Mining Services in
mobile. Data Mining
Services are useful in several sectors including Mobile, WWW,
HealthCare

scenario etc.

We discuss several aspects step
-
by
-
step in
this paper and analyze those aspects and approaches sequentially.
Discuss

their advantages and disadvantages and conclude with new
concept.


The remaining of this paper is organized as follows. We

discuss
CLDC

and J2Me

in Section 2. In Section 3 we discuss about MIDP
.

The
proposed data mining algorithm, namely
CSUA

in section
4
.

In

section
5
we discuss about the
C
hallenges.

The conclusions and
future
directions

are given in Section
6
.

Finally references are given

in section 7
.


2.

CLDC

and J2ME

The J2ME architecture is described in general before the components
in the J2ME technol
ogy are introduced.J2ME applications are also
discussed in general, and it is explained how they are made available
to end users.J2ME is a highly optimized Java runtime environment.
J2ME is aimed at the consumer and embedded devices market. This
includes d
evices such as cellular

telephones, Personal Digital
Assistants (PDAs) and other small
devices. Fig 1

shows the J2ME
architecture. Java 2 Standard Edition (J2SE) developers

should be
familiar with Java Virtual Machines (JVMs) and at least one host
operatin
g
System (OS).


Fig1 J2ME

Architecture

International Journal of Computer Applications (0975


8887)

Volume 7


No.
4
, September 2010

24


The fundamental branches of the J2ME plstform are
configurations
.
A configuration is a specification that describes a Java Virtual
Machine and some set of APIs that are targeted at a specific class of
device.


Th
e Connected, Limited Device Configuration is one such
specification. The CLDC specifies the APIs for devices with less than
512 KB of RAM available for the Java system and an intermittent
(limited) network connection. It specifies a stripped
-
down Java
virt
ual machine, called the KVM, as well as several APIs for
fundamental application services. Three packages are minimalist
versions of the J2SE java.lang, java.io, and java.util packages. A
fourth package, javax.microedition.io, implements the Generic
Connec
tion Framework, a generalized API for making network
connections.


Many J2ME games already exist and enjoy great popularity
especially among young generation. Java

comes

with the immense
requirement of the

object
-
oriented

programming language for
develope
rs to implement new

mobile applications [
7
]. Configurations
provide core functionality

and a way to provide greater flexibility

but
no services

for managing the application life
-
cycle, for driving the
user

interface, for maintaining and updating persistent

data on

the
device or for secure access to information stored on a

network server
[
8
].Fig 2 shows the CLDC position in J2ME Architecture.


Several networks have conducted a survey on users’ watching
behavior [9] which reflects that user behavior pattern r
ecognition is
not so easy task; we can achieve this by CLDC and MIDP
component. Instead of replacing existing TV service, mobile services
should be complementary [10], and offer more interactive means for
users

to watch their chosen content.

The CLDC compo
nent specifies
the connection between the MIDP profile and the connecting
components with the server.


All PDAs are small computing devices that contain an operating
system, processor, memory and a port to connect the PDA to
peripherals and external compu
ting devices.


Fig 2 CLDC in J2ME Architecture




3.

MIDP

The Mobile Information Device Profile is a specification for a J2ME
profile. It is layered on top of CLDC and adds APIs for application
life cycle, user interface, networking, and persistent storage
[11]. An
application written for MIDP is called a
MIDlet
. MIDlet applications
are subclasses of the javax.microedition.midlet.MIDlet class that is
defined by MIDP.

MIDlets are packaged and distributed as
MIDlet suites
. A MIDlet
suite can contain one or mo
re MIDlets. The MIDlet suite consists of
two files:


Java Application Descriptor (.jad) file

The Java Application Descriptor file lists the archive file
name, the names and class names for each MIDlet in the
suite, and other information. This file is used

by the mobile
device to ensure that device has the minimum requirements to
run the application.


A Java Archive file (.jar) file.

The archive file contains the MIDlet classes and resource
files

Java Studio Mobility is an integrated development environmen
t
(IDE), based on the
Net Beans

development platform, that enables
you to use J2ME technologies and add special tools that enable you
to code and test J2ME applications, such as emulators and
obfuscators.

A MIDlet is a J2ME application designed to operat
e on an small
computing device. A MIDlet is defined with at least a single class
that is derived from the javax.micoedition.midlet.MIDlet abstract
class.
The Position of MIDP is shown in Fig 3.




Fig 3 MIDP in J2ME Architectur
e

4.

Proposed Method: CSUA

In this section, we describe the proposed method. The entire novel
concept is divided into
five

phases: 1) Create the Data Set ac
cording
to the user requirement
.2) Select the data set when ever needed. 3)
Update the data set whenev
er needed. 4) Alter the data set.
5) Mobile
communication for the data set.



Algorithm for Creating the Data Set

International Journal of Computer Applications (0975


8887)

Volume 7


No.
4
, September 2010

25



For creating the data set we consider the following things:

a)

Table size

b)

No of Columns

c)

No of rows


Assumptions:


Col: Column

Dt: Data Type

V1
, v2….vn: Values


Algorithm Create Datasets (CDS)


1.

Select the Data Source (WWW, XML, Data Warehouse
etc.)
.

2.

Apply the Create table statement on Data Source

Create table tablename as select * / Col1, Col2, Col n from
data source.

3.

Apply the Create table stat
ement[Stand alone]

Create table tablename (col1 Dt1, Col2 Dt2………Col n
Dtn) with condition clause.

4.

Insert the values considering the domain
.

5.

For Stand alone

Insert into tablename values
(v1, v2
………vn)
.

6.

For step no2 When using universal false condition

Inse
rt into table (Select query)
.

7.

Finish.




Algorithm for selecting the Data Set


For selecting the data set we use select statement


Algorithm Select Datasets (SDS)


1.

Select the Data Source (WWW, XML, Data Warehouse
etc.).

2.

For valuating all columns with rows

S
elect * from datasource.

3.

For selected set of columns

Select col1, col2,…………….coln from datasource.

4.

For some specified set of conditions

Select * / col1
, col2
,……………
..coln from
datasource
where condition.


5. Finish.



Algorithm for Updating the D
ata Set
(UDS)


The update statement is used to change or modify data sets in a
database.


1.

Select the Data Source (WWW, XML, Data Warehouse
etc.) for updating.

2.

U
pdate all the rows from a table

Update

tablename set col1
= v1, col2
=v2………
.coln=vn.

3.

Update select
ed set of rows

Update table set col1=v1
, col2
=v2…….coln=vn where
condition.

4.

Finish
.



Algorithm for Alter the Data Set

(ADS)


By alter statement we can add a column, delete a column and
increase and decrease the size of the column in the database.


1.

Select

the Data Source (WWW, XML, Data Warehouse
etc.) for alter.

2.

Add new column to the data source
.

Alter table tablename add (Col1 Dt1, Col2 Dt2……..Coln
Dtn).

3.

Modify the size of the column

If(table is empty)

{

3a. Increase the size

Alter table tablename modif
y (Col1 Dt1, Col2
Dt2……..Coln Dtn).

3b. Decrease the size

Alter table tablename modify (Col1 Dt1, Col2
Dt2……..Coln Dtn).

}

If(table is not empty)

{

3c. Increase the size

Alter table tablename modify (Col1 Dt1, Col2
Dt2……..Coln Dtn).

}

Else

{

Exit (
0);

}

4.

Dr
op the column

Alter table tablename drop (columns).

5.

Finish



Algorithm for Mobile Communication
for


the Data Set (MCDS)


1.

Select the Data Source (WWW, XML, Data Warehouse
etc.) for alter.

2.

Apply the MIDP Profile.

3.

Using MIDlet establish the connection

4.

Apply the packages of J2ME according to the need.

5.

Using WTK (Wireless ToolKit) access the result on the
simulator.

6.

Apply data Mining techniques.

7.

Finish.


After analyzing the several aspects of CSUA method the picture is
clear for any databases it is easy

to manage and whenever necessary
we can update the repository system.


We also apply several data mining techniques very smoothly because
our data base is consistent because of limiting redundancy in the
database. Finally apply the J2ME for mobile devices

so that we can
coherent the entire above scenario for mobile computing
environments.


5.

CHALLENGES

A number of constraints and technical difficulties faced by
researchers, which are discussed in this section. These general
problems must be considered for fu
rther research in this area to
International Journal of Computer Applications (0975


8887)

Volume 7


No.
4
, September 2010

26


propose new technologies for making mobile computing easier. Some
of these are:


The screen size of the mobile is a big limitation. The screen size
can affect the approximate visualization of complex results
representing the d
iscovered model.


Mobile navigation facility is also a big task to achieve and
implement.


The overhead due to the communication between MIDLET and
Data Mining service should not affect the execution time.


The experiments on system performance depend almost

entirely on
the computing power of the server on which data mining task is
executed.

We are attempting to implement knowledge discovery applications.
Techniques and tools can also be implemented in
DMS

as
decentralized and interoperable services that enab
le the development
of complex system such as distributed knowledge discovery suits.


6.

CONCLUSIONS AND FUTURE
DIRECTIONS

Along with the rapid development of information technology,
executing advanced technologies through mobile handset is the prime
directio
n of development. Implementation of intelligent modules on
mobile devices through the combination of J2ME and
related

computing will be the base to introduce data mining features in
Mobile Computing.
In future we propose a novel data mining
algorithm named

J2ME
-
based Mobile Progressive Pattern Mine
(J2MPP
-
Mine) for effective mobile computing. In J2MPP
-
Mine, we
first propose a subset finder strategy named Subset
-
Finder (S
-
Finder)
to find the possible subsets for prune. Then, we propose a Subset
pruner algori
thm (SB
-
Pruner) for determining the frequent pattern.
Furthermore, we proposed the novel prediction strategy to determine
the superset and remove the subset which generates a less number of
sets due to different filtering pruning strategy which is more eff
icient
and less time consuming.














7.

REFERENCES

[1]
M. Migliardi, M. Maheswaran, B. Maniymaran, and P. Mobile

Interfaces to Computational, Data, and Service Grid

Systems.
ACM
SIGMOBILE Mobile Computing and Communications

Review
, 6(4),
2004
.

[2]
S. Wesner, T. Dimitrakos, and K. Jeffrey. Akogrimo


The

Grid
goes Mobile.
ERCIM
,
October (
59), 2004.

[3]
A. Arcelus, M.H. Jones, R. Goubran, and F. Knoefel. Advanced

Information Networking and

Applications Workshops, 2007,

AINAW.
21st International

Conference,

May 2007.

[4]
Raphael M. Bahati and Michael A. Bauer. In

ICAS ’08:

Proceedings of the Fourth International Conference

on
Autonomic
and Autonomous Systems (ICAS’08), pages 88

93, Washington, DC,
USA, 2008. IEEE Computer Society.

[5

Michael Beetz, Jan Bandouch, Alexandra Kirsch, Alexis

Maldonado, Armin M¨uller, and Radu Bogdan Rusu. In

Proceedings
of the 4th COE Workshop

on Human Adaptive
Mechatronics
(HAM), 2007.

[6]
Ralph Bergma
nn. Ambient intelligence for decision making

in
fire service organizations. In AmI, pages 73

90, 2007.

[7]
Isakow, A. and Shi, H. “Review of J2ME and J2MEbased

Mobile
Applications”, International Journal of

Communication
and Network
Secur
ity, Vol. 8 No. 2,pp. 189
-
198.

[8]
Ortiz, 2004a, A Survey of J2ME Today, Sun

Developer Network

(SDN), viewed 13 August 2
007.

[9]
http://www.3g.co.uk/PR/Sept2005/1943.htm

[10
]
http://www.cellular
-
news.com/story/18707.php

[1
1
]
http://docs.sun.com
.